Open Loop Jump Observed 02 February 2009

SUMMARY

The open loop jump observed was not caused by any issue in the PCS tracking polynomials.

Dave Thompson reported the following:

I could use some help tracing out exactly what happened Sunday night (UT Date 20090202) when, during open-loop observations, we experienced a 1.7 arcsec jump that jumped back about 110 seconds later. The attached plot shows the radial offset from the median centroid XY vs. UT time from the guider images:

• jump.pdf: jump.pdf

and a listing of the data right around the jump is given below. The jumps are obvious in the XY centroids. Note: the guide corrections were turned off via the PCSGUI so the actual guide corrections were not being applied to the mount.

Guider Image Name Xcent Ycent UT Time TELALT EXPTIME
guiderimage004162 56.473 47.447 03:45:05.97 61.010 4.6
guiderimage004163 51.836 90.194 03:45:33.97 61.109 4.6
guiderimage004164 33.926 74.954 03:45:40.72 61.133 4.6
guiderimage004165 26.991 71.202 03:45:47.25 61.156 4.6
guiderimage004166 29.305 70.800 03:45:53.88 61.179 4.6
guiderimage004167 28.588 76.822 03:46:00.74 61.203 4.6
guiderimage004168 26.945 71.391 03:46:07.13 61.225 4.6
guiderimage004169 31.700 71.675 03:46:13.94 61.249 4.6
guiderimage004170 32.963 71.460 03:46:20.60 61.273 4.6
guiderimage004171 29.029 73.420 03:46:26.66 61.294 4.6
guiderimage004172 25.548 72.990 03:46:33.60 61.317 4.6
guiderimage004173 29.362 70.951 03:46:40.18 61.341 4.6
guiderimage004174 29.334 71.052 03:46:46.72 61.364 4.6
guiderimage004175 25.833 72.900 03:46:53.16 61.386 4.6
guiderimage004176 28.943 68.107 03:46:59.93 61.410 4.6
guiderimage004177 30.242 71.883 03:47:06.40 61.433 4.6
guiderimage004178 29.055 75.410 03:47:13.18 61.457 4.6
guiderimage004179 31.331 73.991 03:47:20.19 61.481 4.6
guiderimage004180 49.096 46.729 03:47:26.27 61.503 4.6

The relevant time ranges to investigate are:
03:45:05.97 to 03:45:33.97 the first jump occurred, immediately followed by another in
03:45:33.97 to 03:45:40.72 where the telescope remained for about 110 seconds, then
03:47:20.19 to 03:47:26.27 when the image jumped back

Michele's Analysis

First, I can say this is not the "T0" problem at the PCS level. The "T0" problem is an issue observed in the low-level mount data (source unknown at this time) where the T0 values (epoch of the tracking polynomial) backtracked in time and then proceeded in a normal fashion. I have not seen this problem in the T0 values that I have examined which PCS records.

I then examined the azimuth and elevation trajectories in Matlab. The X-axis in these plots is just an Index which is equivalent to the PCS update instance. I can see the radial offset for the time of concern \xB1 10 seconds is on the order of ~11 milliarcseconds. This is well below the 1.7" Dave has reported for the jump. Interestingly, I can also see a periodicity in the data which occurs every 7-8 polynomial updates (~0.4 seconds); the cause of this periodicity is unknown. There is no explicit update in PCS for this period. I will be contacting Dave Terrett to see if he can shed some light on this. The order of magntitude of this periodicity is the ~11 mas quoted above.

• radialOffset.jpg:
  

I will note that I can see in the elevation plot, the 2 second sky patch update (~2.4 mas) performed by PCS; these are the values dipping to -2.4.

• elevationOffset.jpg:
  

For completeness, the azimuth plot is also attached. The 2 second update is only occasionally apparent.

• azimuthOffset.jpg:
  

-- MicheleDeLaPena - 09 Feb 2009